The authors linked these measurements to clinical and radiological characteristics and found that markers of neuronal or astroglial injury correlated with poor outcome, in particular when they increased over time. Inflammatory markers, which were lower in ventricular than lumbar CSF, did not predict outcome.
This very interesting report leaves the readers with a few questions relating to its analysis. What is the correlation structure between serum and lumbar and ventricular CSF for the different markers?
The markers as presented in Table 3 of the manuscript [ 1 ], were they measured in lumbar or ventricular CSF samples? Rohlwink et al found that CSF neuromarkers are predictive of mortality in their pediatric cohort. Also in adults, NSE levels decrease over time to a median Neuron-specific enolase as predictor for day mortality in adult HIV-negative tuberculous meningitis patients.
Abbreviation: NSE, neuron-specific enolase. We think that inflammatory markers should not be disregarded in pediatric TBM. Although the authors measured relevant markers, they did not find a relation with outcome, possibly because of inadequate power. This is in contrast with findings in adult HIV-negative TBM, which have shown lower Th1 and Th2 cytokines Vietnam, patients [ 3 ] and higher temperature and CSF neutrophil proportions Indonesia, patients [ 2 ] to be associated with a higher mortality.
Moreover, reported benefits of anti-tumor necrosis factor and other anti-inflammatory agents for paradoxical worsening of TBM suggest that inflammation is important in the pathogenesis of TBM [ 4 ]. When numbers allow, it would also be interesting to look into genetic determinants of inflammatory response and brain damage. In a recent study, using an unbiased approach measuring metabolites, we identified low CSF but not serum tryptophan as predictive for better survival.
